From makeup to memories: AI opens new possibilities for blind users

Artificial intelligence tools are increasingly acting as a "textual mirror" for blind people, offering descriptions of their bodies and surroundings that were previously inaccessible and reshaping how they engage with the visual world.

Applications such as Be My Eyes, Envision and Aira Explorer use AI to describe facial features, clothing and environments, helping users manage personal grooming, navigate daily tasks and access visual memories. For some blind users, the technology provides a way to form an understanding of their own appearance, particularly for those who have never seen themselves or who lost their sight later in life, says the BBC.

"All our lives, blind people have had to grapple with the idea that seeing ourselves is impossible," said Lucy Edwards, a blind content creator. "Suddenly we have access to all this information about ourselves, about the world, it changes our lives."

Edwards, who lost her sight after childhood, said the technology filled a long-standing gap in self-perception. "I had sight for 17 years of my life, and while I could always ask people to describe things to me, the truth is that I haven't had an opinion about my face for 12 years," she said.

Users employ these tools for activities ranging from skincare and makeup routines to coordinating outfits. Beyond appearance, the applications assist with practical tasks such as shopping, reading letters and describing old photographs.

"Often the first question they ask is how they look," said Karthik Mahadevan, chief executive of Envision.

Edwards said AI descriptions have allowed her to access visual memories that were once closed off. "Suddenly AI can describe every photo on the internet and it can even tell me what I looked like next to my husband on my wedding day," she said.

Some users also experiment with appearance-based feedback. "Suddenly I'm taking a photo and I can ask AI to give me all the details, to give me a score out of 10, and although it's not the same as seeing, it's the closest I'm going to get for now," Edwards said.

Advocates say the technology offers greater independence, but researchers and users warn that it can also introduce emotional and psychological risks, particularly around body image.

"We have seen that people who seek more feedback about their bodies, in all areas, have lower body image satisfaction. AI is opening up this possibility for blind people," said Helena Lewis-Smith, a health psychology researcher.

She said AI feedback can reinforce unrealistic ideals. "AI's processing can return a photo with a lot of changes that make the person look totally different, implying that all of this is what they should change, and therefore that the way they look now is not good enough," Lewis-Smith said.

"What is scary now is that AI not only allows blind people to do this by comparing themselves to descriptions of photos of other human beings, but also to what AI might consider the perfect version of them," she added.

Edwards said the technology can amplify insecurities when users seek critical feedback. "These descriptions have the potential to change the way we feel about ourselves. But this can also be used in a negative way, because maybe you don't like something about yourself, and you tell the AI that you're not sure about a feature of your body... it may also tell you, 'You're right, here's how you can change it'," she said.

Examples of direct AI feedback shared with users include statements such as, "Your skin is hydrated, but it definitely doesn't look like the almost perfect example of reflective skin, with non-existent pores as if it were glass, in beauty ads," and, "Maybe if your jaw was less elongated (...) your face would look a little more like what is objectively considered beautiful in your culture."

Lewis-Smith said such pressure can have broader consequences. "The more pressure people have about their bodies, the more cases of mental health conditions such as depression and anxiety increase," she said.

Technical limitations further complicate the use of AI as a visual aid. Mahadevan said Envision identified accuracy problems as its tools developed. "At first, the descriptions were very good, but we noticed that many of them were inaccurate and changed important details, or invented information when what was in the image didn't seem to be enough," he said.

Blind AI user Joaquín Valentinuzzi said errors can undermine trust. "Sometimes it changed my hair colour or described my expressions incorrectly... This kind of thing can make you feel insecure, especially if... we trust these tools and use them as a way to gain self-knowledge," he said.

Mahadevan said user control is a central design feature. "People being able to control the information they receive is one of the main features of our products, because AI can learn their preferences and desires and give people the information they need to hear," he said.

Researchers note that AI systems lack an understanding of context and individuality. "Body image is not one-dimensional and is made up of several factors, such as context, the type of people we want to compare ourselves to, and the things we are capable of doing with our bodies... All of this is something that AI does not understand," said Meryl Alper, a researcher at Northeastern University.

Street, a cited expert, said current systems remain limited. "AI today can tell you that you have a sideways smile. But for now it can't analyse all your photos and tell you that, for example, you have the same expression as when you were enjoying the Sun on the beach," they said.

Alper said the field remains under-studied. "All of this is in its infancy, and there really isn't any kind of massive research on the effect of these technologies, with their biases, errors, and imperfections, on the lives of blind people," she said.

As the technology continues to evolve, many blind users describe access to visual information as transformative, while acknowledging that its emotional and psychological impact is something they are still learning to navigate.